On Tue, Feb 27, 2024 at 8:53 PM Dean Rasheed <dean.a.rash...@gmail.com> wrote:
>
> Attached is a very rough patch. It seemed better to build the
> projection in the executor rather than the planner, since then the
> extra work can be avoided, if EPQ is not invoked.
>
> It seems to work (it passes the isolation tests, and I couldn't break
> it in ad hoc testing), but it definitely needs tidying up, and it's
> hard to be sure that it's not overlooking something.
>
Hi. minor issues.
In nodeAppend.c, nodeMergeAppend.c
+ if (estate->es_epq_active != NULL)
+ {
+ /*
+ * We are inside an EvalPlanQual recheck. If there is a relevant
+ * rowmark for the append relation, return the test tuple if one is
+ * available.
+ */
+ oldcontext = MemoryContextSwitchTo(estate->es_query_cxt);
+
+ node->as_epq_tupdesc = lookup_rowtype_tupdesc_copy(tupType, tupTypmod);
+
+ ExecAssignExprContext(estate, &node->ps);
+
+ node->ps.ps_ProjInfo =
+ ExecBuildProjectionInfo(castNode(Append, node->ps.plan)->epq_targetlist,
+ node->ps.ps_ExprContext,
+ node->ps.ps_ResultTupleSlot,
+ &node->ps,
+ NULL);
+
+ MemoryContextSwitchTo(oldcontext);
EvalPlanQualStart, EvalPlanQualNext will switch the memory context to
es_query_cxt.
so the memory context switch here is not necessary?
do you think it's sane to change
`ExecAssignExprContext(estate, &node->ps);`
to
`
/* need an expression context to do the projection */
if (node->ps.ps_ExprContext == NULL)
ExecAssignExprContext(estate, &node->ps);
`
?
